Methods, computer program products and/or systems are provided that perform the following operations: obtaining a performance matrix representing accuracies obtained by executing a plurality of pipelines on a plurality of training data sets, wherein a pipeline comprises a series of operations performed on a data set; selecting a defined number of top pipelines as potential pipelines for a testing data set based, at least in part, on a similarity between the testing data set and each of the plurality of training data sets represented in the performance matrix; storing results from executing each of the potential pipelines as a new data set; determining a pipeline accuracy for each of the potential pipelines when executed against the testing data set; and providing a recommended pipeline for use with the testing data set based, at least in part, on the pipeline accuracy for each potential pipeline.
Legal claims defining the scope of protection. Each claim is shown in both the original legal language and a plain English translation.
4. The computer-implemented method of claim 2, wherein when computing the similarity between the testing data set and each of the plurality of training data sets, only entries that are present in both the training data set and the testing data set being compared for similarity are used.
This computer-implemented method automates machine learning by first obtaining a performance matrix, which stores the accuracies of various machine learning pipelines when run on multiple training datasets. A pipeline is a sequence of operations applied to data. Next, a set of top-performing pipelines is selected as potential candidates for a specific testing dataset. This selection is based on how similar the testing dataset is to each of the training datasets represented in the performance matrix. Crucially, when calculating this similarity between a testing dataset and a training dataset, **only the data points or features that exist in *both* datasets being compared are used for the similarity calculation.** After selection, the method stores the execution results of these potential pipelines, determines their accuracy against the testing dataset, and finally recommends the best pipeline for that testing dataset.
7. The computer-implemented method of claim 1, wherein each (i, j) entry in the performance matrix represents an accuracy obtained by executing a pipeline i and a training data set j.
This computer-implemented method automates machine learning by first generating and obtaining a performance matrix. This matrix is central to the process, storing the performance metrics of various machine learning pipelines. Specifically, **each entry at position (i, j) within this performance matrix represents the accuracy score achieved when a particular machine learning pipeline 'i' was executed on a specific training dataset 'j'.** Following this, the method selects potential top pipelines for a new testing dataset, basing the choice on the similarity between the testing dataset and the training datasets already in the matrix. It then stores the results of executing these potential pipelines, evaluates their accuracy against the testing dataset, and ultimately provides a recommended pipeline for optimal use with that testing data.
Cooperative Patent Classification codes for this invention. Click any code to explore related patents in that topic.
August 10, 2020
March 26, 2024
Browse 5M+ US patents with plain-English claim translations and AI-generated analysis.